The most successful 21st century businesses were born digital. Amazon changed retail by moving data so efficiently that packages arrive at your door the next day. Uber changed the taxi business by having algorithms replace dispatchers. Netflix changed entertainment by making media available for immediate viewing.
Businesses born in the previous century, in turn, must adapt to survive. They have to take workflows established decades ago and transform them to meet the demands of the modern consumer.
Most of the data embedded in enterprise operations is still largely unstructured, and thus limits true progress towards digitalization. The fundamental vehicle for exchanging information is the document, which is formatted for people. However, to operate at faster-than-human-speeds, where decisions are made by algorithms and operations move at machine speeds, the data in documents must be extracted and re-formatted for APIs.
Automation unquestionably plays a key role in updating operations. But it can only succeed—and offer scalability—if the data it is fed has been cleaned up and formatted for digital consumption. The fundamental problem is that automation and workflow solutions are designed to act on data, not read it. So before data can power an automated process or drive an algorithmic decision or fuel business intelligence, it must first be extracted from documents and converted to machine-readable formats, like JSON or XML.
Deloitte’s 2021 global intelligent automation survey reaffirms a trend that has been prevalent since the early days of RPA: 73% of the executives who responded have made investments in automation but only 13% of them have achieved automation at scale.
COVID-19 has further exposed operational inefficiencies. When the human workforce could not go to the office, the lack of end-to-end process integration could no longer be ignored. Companies had to confront their paper-based processes and think beyond merely outsourcing to lower-cost sources of labor. Sure, Global Business Services offer domain expertise and lower cost labor markets, but processes remain manual, transaction times remain stuck at human speed, and this workforce, too, is affected by a global pandemic.
The key to survival is to provide a frictionless experience to customers, starting with their first point of interaction with the business. What prevents this from happening is that unstructured data—including contracts, forms, and emails—is in the wrong format for direct processing by an API or algorithm. This means that human staff must manually process this information and key it into downstream systems. The goal of any digital transformation is to improve the flow of data through an enterprise. And it starts with getting the data in documents into digital formats so that it can be processed at machine speeds.
Attempts to digitize documents have been hampered for years by the limitations of rule-based approaches that relied on documents to be formatted consistently. Newer approaches leaned on natural language processing and were not bound by rules, but they, too, have revealed limitations of their own. These methods consider the sequential meaning of words but cannot understand the meaning conveyed by formatting features, such as tables and checkboxes. Technology that focuses on textual semantics works well on documents with paragraphs of text, like a contract, but fails when there are important visual cues in the formatting that tell the reader how to interpret the information.
In contrast to the competition, Applica combines technologies that previously processed only structured or only unstructured documents to create an innovative hybrid model that can automate a vast range of document types, whether scanned from paper or born-digital. Our proprietary language model, TILT (Text-Image-Layout Transformer), simultaneously learns layout, visual aspects, and textual semantics, along with exhibiting an unprecedented ability to “read” graphical embeddings.
Today Applica is the solution that’s taking workflows into the 21st century. Business today runs on APIs, and Applica converts data into API formats. We make it possible for banks to now open new accounts in minutes rather than days. We make it possible for insurance companies to make underwriting decisions and settle claims in hours rather than weeks. We make it possible for mortgage lenders to close a loan in weeks rather than months. We make it possible for clinical labs to process hundreds of thousands of test requests in a single day. In other words, our tech enables the $1.5 billion digital transformation market by solving the unstructured text problem for business.